One Visit, Four Diabetes Codes: What Can Actually Be Submitted
An established 67-year-old Medicare Advantage member is seen by her primary care physician on August 14, 2025. The assessment lists type 2 diabetes with diabetic chronic kidney disease, stage 3b (E11.22, N18.32), long-term insulin use, and a BMI of 38.4. The note also states: "Diabetic retinopathy, mild NPDR, stable; ophthalmology managing, last dilated exam 11 months ago." The problem list carries three more entries: "Diabetic foot ulcer, resolved 2022," "Amputation, right BKA 2021, ambulates with prosthesis," and "Uncontrolled diabetes." No glucose meter readings, no hypoglycemic or hyperglycemic episodes, and no foot exam findings appear in this note; the A1c is 9.1%. The prior-year claim submitted E11.9, N18.9, and Z89.511. A payer chart-review vendor has flagged the record for review, and the coder has to decide what is supportable from this encounter and what needs follow-up before the chart is closed.
1. Which diagnosis codes are supportable from this encounter, and which are not? Walk through each candidate and justify your decision.
Read the worked answer
Supportable
- E11.22 + N18.32. The note links the diabetes to the CKD and names the stage. E11.22 is a combination code that requires an additional code for the CKD stage, so both appear. These land in two different hierarchies, so both can be captured.
- Z79.4 (long-term insulin use) if insulin use is confirmed as long-term. It supports completeness but does not itself drive payment.
- Z68.38 as a secondary code for the adult BMI of 38.4. BMI codes are never first-listed and are not a payment driver; they are supporting data.
Not supportable as submitted
- E11.9. The note documents complications, so the uncomplicated code contradicts the record. Under the current V28 model, uncomplicated type 2 diabetes does not map to a payment HCC anyway; carrying it forward adds risk without revenue.
- N18.9. Unspecified CKD is both superseded by the staged code and, in the current model, not a risk-adjusting category. Submitting it is a red flag to auditors that the coder worked from the problem list rather than the note.
- "Uncontrolled diabetes." There is no ICD-10-CM code for this phrase. An A1c of 9.1% does not by itself establish hyperglycemia; E11.65 requires the provider to document (and act on) hyperglycemia.
- Resolved foot ulcer. Not active disease in 2025; it cannot be carried forward.
- Z89.511 is codeable because amputation status is documented, but do not assume it pays. Verify the current model's mapping file rather than the code title before treating it as an HCC.
2. Explain how hierarchy rules change the outcome when several codes in the same disease family are documented, and why the diabetes-with-CKD code and the CKD stage code can both be captured.
Read the worked answer
A hierarchy is a ranking within one disease family: only the most severe documented category in that family is paid for the year. If the record supported both E11.9 and E11.22, the coder still submits the most specific, clinically accurate code - and the model pays one diabetes category, the higher one. Two practical consequences follow.
First, a hierarchy is not a reason to submit everything and let the model sort it out. Unsupported codes are audit findings even when the hierarchy would have suppressed them, because the submitted diagnosis data itself must be accurate.
Second, hierarchies are family-specific, not body-system-wide. Diabetes with CKD sits in the diabetes family; CKD stage sits in the renal family. Because they are separate hierarchies, capturing both is correct and additive - but only if the documentation genuinely establishes both. That is why E11.22 requires the additional N18.x code: the diabetes code does not carry the stage, and the stage code does not carry the diabetes linkage.
When a single combination code could describe two things, check whether an additional code is required or excluded before deciding the pair is redundant. Here it is required.
3. What would you do about the retinopathy statement, the resolved foot ulcer, the phrase "uncontrolled diabetes," and the amputation status? Distinguish documentation/query issues from mapping issues.
Read the worked answer
Retinopathy - documentation issue first. "Stable, ophthalmology managing" is a referral mention, not an assessment of the retinopathy in this encounter, and it does not establish type, severity, or laterality, which the combination code requires. The compliant path is to obtain the ophthalmology note from a face-to-face encounter in the same data year and capture the diagnosis from that source, or send a non-leading query to the PCP asking whether the retinopathy was evaluated at this visit and, if so, its current status. Do not infer the severity from "mild NPDR" in a problem list that was not addressed.
Resolved foot ulcer - pure coding rule. No query needed. A resolved condition is not active disease; it cannot be reported for this year.
"Uncontrolled diabetes" - query issue. The term has no code. If the provider believes the patient is hyperglycemic, the query should ask whether the diabetes is currently associated with hyperglycemia and whether it was managed at this visit. The A1c alone does not answer that. Absent a provider response, code nothing beyond what is documented.
Amputation status - mapping issue. Z89.511 is correct coding if status is documented, but payment depends entirely on the model's HCC mapping file, which changes between model versions. Coders should check the mapping, not the code descriptor, and should never carry a status code forward on the assumption that it pays.